models.py

from datetime import datetime, timedelta
from enum import Enum, IntEnum
from flask import current_app, url_for
from flask_hashids import HashidMixin
from flask_login import UserMixin
from sqlalchemy.ext.associationproxy import association_proxy
from time import sleep
from tqdm import tqdm
from werkzeug.security import generate_password_hash, check_password_hash
from werkzeug.utils import secure_filename
import json
import jwt
import os
import requests
import secrets
import shutil
import xml.etree.ElementTree as ET
import yaml
from app import db, hashids, login, mail, socketio
from app.converters.vrt import normalize_vrt_file
from app.email import create_message


##############################################################################
# enums                                                                      #
##############################################################################
# region enums
class CorpusStatus(IntEnum):
    UNPREPARED = 1
    SUBMITTED = 2
    QUEUED = 3
    BUILDING = 4
    BUILT = 5
    FAILED = 6
    STARTING_ANALYSIS_SESSION = 7
    RUNNING_ANALYSIS_SESSION = 8
    CANCELING_ANALYSIS_SESSION = 9


class JobStatus(IntEnum):
    INITIALIZING = 1
    SUBMITTED = 2
    QUEUED = 3
    RUNNING = 4
    CANCELING = 5
    CANCELED = 6
    COMPLETED = 7
    FAILED = 8


class Permission(IntEnum):
    '''
    Defines User permissions as integers by the power of 2. User permission
    can be evaluated using the bitwise operator &.
    '''
    ADMINISTRATE = 1
    CONTRIBUTE = 2
    USE_API = 4


class UserSettingJobStatusMailNotificationLevel(IntEnum):
    NONE = 1
    END = 2
    ALL = 3


class ProfilePrivacySettings(IntEnum):
    SHOW_EMAIL = 1
    SHOW_LAST_SEEN = 2
    SHOW_MEMBER_SINCE = 4


class CorpusFollowerPermission(IntEnum):
    VIEW = 1
    CONTRIBUTE = 2
    ADMINISTRATE = 4
# endregion enums


##############################################################################
# mixins                                                                     #
##############################################################################
# region mixins
class FileMixin:
    '''
    Mixin for db.Model classes. All file related models should use this.
    '''
    creation_date = db.Column(db.DateTime, default=datetime.utcnow)
    filename = db.Column(db.String(255))
    mimetype = db.Column(db.String(255))

    def file_mixin_to_json_serializeable(self, backrefs=False, relationships=False):
        return {
            'creation_date': f'{self.creation_date.isoformat()}Z',
            'filename': self.filename,
            'mimetype': self.mimetype
        }
    
    @classmethod
    def create(cls, file_storage, **kwargs):
        filename = kwargs.pop('filename', file_storage.filename)
        mimetype = kwargs.pop('mimetype', file_storage.mimetype)
        obj = cls(
            filename=secure_filename(filename),
            mimetype=mimetype,
            **kwargs
        )
        db.session.add(obj)
        db.session.flush(objects=[obj])
        db.session.refresh(obj)
        try:
            file_storage.save(obj.path)
        except (AttributeError, OSError) as e:
            current_app.logger.error(e)
            db.session.rollback()
            raise e
        return obj
# endregion mixins


##############################################################################
# type_decorators                                                            #
##############################################################################
# region type_decorators
class IntEnumColumn(db.TypeDecorator):
    impl = db.Integer

    def __init__(self, enum_type, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self.enum_type = enum_type

    def process_bind_param(self, value, dialect):
        if isinstance(value, self.enum_type) and isinstance(value.value, int):
            return value.value
        elif isinstance(value, int):
            return self.enum_type(value).value
        elif isinstance(value, str):
            return self.enum_type[value].value
        else:
            return TypeError()

    def process_result_value(self, value, dialect):
        return self.enum_type(value)


class ContainerColumn(db.TypeDecorator):
    impl = db.String

    def __init__(self, container_type, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self.container_type = container_type

    def process_bind_param(self, value, dialect):
        if isinstance(value, self.container_type):
            return json.dumps(value)
        elif isinstance(value, str) and isinstance(json.loads(value), self.container_type):
            return value
        else:
            return TypeError()

    def process_result_value(self, value, dialect):
        return json.loads(value)
# endregion type_decorators


##############################################################################
# Models                                                                     #
##############################################################################
# region models
class Role(HashidMixin, db.Model):
    __tablename__ = 'roles'
    # Primary key
    id = db.Column(db.Integer, primary_key=True)
    # Fields
    name = db.Column(db.String(64), unique=True)
    default = db.Column(db.Boolean, default=False, index=True)
    permissions = db.Column(db.Integer, default=0)
    # Relationships
    users = db.relationship('User', back_populates='role', lazy='dynamic')

    def __repr__(self):
        return f'<Role {self.name}>'

    def add_permission(self, permission):
        if not self.has_permission(permission):
            self.permissions += permission

    def has_permission(self, permission):
        return self.permissions & permission == permission

    def remove_permission(self, permission):
        if self.has_permission(permission):
            self.permissions -= permission

    def reset_permissions(self):
        self.permissions = 0

    def to_json_serializeable(self, backrefs=False, relationships=False):
        json_serializeable = {
            'id': self.hashid,
            'default': self.default,
            'name': self.name,
            'permissions': [
                x.name for x in Permission
                if self.has_permission(x.value)
            ]
        }
        if relationships:
            json_serializeable['users'] = {
                x.hashid: x.to_json_serializeable(relationships=True)
                for x in self.users
            }
        return json_serializeable

    @staticmethod
    def insert_defaults():
        roles = {
            'User': [],
            'API user': [Permission.USE_API],
            'Contributor': [Permission.CONTRIBUTE],
            'Administrator': [
                Permission.ADMINISTRATE,
                Permission.CONTRIBUTE,
                Permission.USE_API
            ],
            'System user': []
        }
        default_role_name = 'User'
        for role_name, permissions in roles.items():
            role = Role.query.filter_by(name=role_name).first()
            if role is None:
                role = Role(name=role_name)
            role.reset_permissions()
            for permission in permissions:
                role.add_permission(permission)
            role.default = role.name == default_role_name
            db.session.add(role)
        db.session.commit()


class Token(db.Model):
    __tablename__ = 'tokens'
    # Primary key
    id = db.Column(db.Integer, primary_key=True)
    # Foreign keys
    user_id = db.Column(db.Integer, db.ForeignKey('users.id'))
    # Fields
    access_token = db.Column(db.String(64), index=True)
    access_expiration = db.Column(db.DateTime)
    refresh_token = db.Column(db.String(64), index=True)
    refresh_expiration = db.Column(db.DateTime)
    # Relationships
    user = db.relationship('User', back_populates='tokens')

    def expire(self):
        self.access_expiration = datetime.utcnow()
        self.refresh_expiration = datetime.utcnow()

    @staticmethod
    def clean():
        """Remove any tokens that have been expired for more than a day."""
        yesterday = datetime.utcnow() - timedelta(days=1)
        Token.query.filter(Token.refresh_expiration < yesterday).delete()


class Avatar(HashidMixin, FileMixin, db.Model):
    __tablename__ = 'avatars'
    # Primary key
    id = db.Column(db.Integer, primary_key=True)
    # Foreign keys
    user_id = db.Column(db.Integer, db.ForeignKey('users.id'))
    # Relationships
    user = db.relationship('User', back_populates='avatar')

    @property
    def path(self):
        return os.path.join(self.user.path, 'avatar')

    def delete(self):
        try:
            os.remove(self.path)
        except OSError as e:
            current_app.logger.error(e)
        db.session.delete(self)

    def to_json_serializeable(self, backrefs=False, relationships=False):
        json_serializeable = {
            'id': self.hashid,
            **self.file_mixin_to_json_serializeable()
        }
        return json_serializeable


class CorpusFollowerAssociation(HashidMixin, db.Model):
    __tablename__ = 'corpus_follower_associations'
    # Primary key
    id = db.Column(db.Integer, primary_key=True)
    # Foreign keys
    corpus_id = db.Column(db.Integer, db.ForeignKey('corpora.id'))
    follower_id = db.Column(db.Integer, db.ForeignKey('users.id'))
    # Fields
    permissions = db.Column(db.Integer, default=0, nullable=False)
    # Relationships
    corpus = db.relationship('Corpus', back_populates='corpus_follower_associations')
    follower = db.relationship('User', back_populates='corpus_follower_associations')

    def __repr__(self):
        return f'<CorpusFollowerAssociation {self.follower.__repr__()} ~ {self.corpus.__repr__()}>'

    def has_permission(self, permission: CorpusFollowerPermission):
        return self.permissions & permission.value == permission.value
    
    def add_permission(self, permission: CorpusFollowerPermission):
        if not self.has_permission(permission):
            self.permissions += permission.value
    
    def remove_permission(self, permission: CorpusFollowerPermission):
        if self.has_permission(permission):
            self.permissions -= permission.value

    def to_json_serializeable(self, backrefs=False, relationships=False):
        json_serializeable = {
            'id': self.hashid,
            'permissions': [
                x.name for x in CorpusFollowerPermission
                if self.has_permission(x)
            ],
            'corpus': self.corpus.to_json_serializeable(),
            'follower': self.follower.to_json_serializeable()
        }
        return json_serializeable


class User(HashidMixin, UserMixin, db.Model):
    __tablename__ = 'users'
    # Primary key
    id = db.Column(db.Integer, primary_key=True)
    # Foreign keys
    role_id = db.Column(db.Integer, db.ForeignKey('roles.id'))
    # Fields
    email = db.Column(db.String(254), index=True, unique=True)
    username = db.Column(db.String(64), index=True, unique=True)
    password_hash = db.Column(db.String(128))
    confirmed = db.Column(db.Boolean, default=False)
    member_since = db.Column(db.DateTime(), default=datetime.utcnow)
    setting_job_status_mail_notification_level = db.Column(
        IntEnumColumn(UserSettingJobStatusMailNotificationLevel),
        default=UserSettingJobStatusMailNotificationLevel.END
    )
    last_seen = db.Column(db.DateTime())
    full_name = db.Column(db.String(64))
    about_me = db.Column(db.String(256))
    location = db.Column(db.String(64))
    website = db.Column(db.String(128))
    organization = db.Column(db.String(128))
    is_public = db.Column(db.Boolean, default=False)
    profile_privacy_settings = db.Column(db.Integer(), default=0)
    # Relationships
    avatar = db.relationship(
        'Avatar',
        back_populates='user',
        cascade='all, delete-orphan',
        uselist=False
    )
    corpora = db.relationship(
        'Corpus',
        back_populates='user',
        cascade='all, delete-orphan',
        lazy='dynamic'
    )
    corpus_follower_associations = db.relationship(
        'CorpusFollowerAssociation',
        back_populates='follower',
        cascade='all, delete-orphan'
    )
    followed_corpora = association_proxy(
        'corpus_follower_associations',
        'corpus',
        creator=lambda c: CorpusFollowerAssociation(corpus=c)
    )
    jobs = db.relationship(
        'Job',
        back_populates='user',
        cascade='all, delete-orphan',
        lazy='dynamic'
    )
    role = db.relationship(
        'Role',
        back_populates='users'
    )
    spacy_nlp_pipeline_models = db.relationship(
        'SpaCyNLPPipelineModel',
        back_populates='user',
        cascade='all, delete-orphan',
        lazy='dynamic'
    )
    tesseract_ocr_pipeline_models = db.relationship(
        'TesseractOCRPipelineModel',
        back_populates='user',
        cascade='all, delete-orphan',
        lazy='dynamic'
    )
    tokens = db.relationship(
        'Token',
        back_populates='user',
        cascade='all, delete-orphan',
        lazy='dynamic'
    )
    
    def __init__(self, **kwargs):
        super().__init__(**kwargs)
        if self.role is not None:
            return
        if self.email == current_app.config['NOPAQUE_ADMIN']:
            self.role = Role.query.filter_by(name='Administrator').first()
        else:
            self.role = Role.query.filter_by(default=True).first()

    def __repr__(self):
        return f'<User {self.username}>'

    @property
    def jsonpatch_path(self):
        return f'/users/{self.hashid}'

    @property
    def password(self):
        raise AttributeError('password is not a readable attribute')

    @password.setter
    def password(self, password):
        self.password_hash = generate_password_hash(password)

    @property
    def path(self):
        return os.path.join(
            current_app.config.get('NOPAQUE_DATA_DIR'), 'users', str(self.id))

    @staticmethod
    def create(**kwargs):
        user = User(**kwargs)
        db.session.add(user)
        db.session.flush(objects=[user])
        db.session.refresh(user)
        try:
            os.mkdir(user.path)
            os.mkdir(os.path.join(user.path, 'spacy_nlp_pipeline_models'))
            os.mkdir(os.path.join(user.path, 'tesseract_ocr_pipeline_models'))
            os.mkdir(os.path.join(user.path, 'corpora'))
            os.mkdir(os.path.join(user.path, 'jobs'))
        except OSError as e:
            current_app.logger.error(e)
            db.session.rollback()
            raise e
        return user

    @staticmethod
    def insert_defaults():
        nopaque_user = User.query.filter_by(username='nopaque').first()
        system_user_role = Role.query.filter_by(name='System user').first()
        if nopaque_user is None:
            nopaque_user = User.create(
                username='nopaque',
                role=system_user_role
            )
            db.session.add(nopaque_user)
        elif nopaque_user.role != system_user_role:
            nopaque_user.role = system_user_role
        db.session.commit()

    @staticmethod
    def reset_password(token, new_password):
        try:
            payload = jwt.decode(
                token,
                current_app.config['SECRET_KEY'],
                algorithms=['HS256'],
                issuer=current_app.config['SERVER_NAME'],
                options={'require': ['exp', 'iat', 'iss', 'purpose', 'sub']}
            )
        except jwt.PyJWTError:
            return False
        if payload.get('purpose') != 'User.reset_password':
            return False
        user_hashid = payload.get('sub')
        user_id = hashids.decode(user_hashid)
        user = User.query.get(user_id)
        if user is None:
            return False
        user.password = new_password
        db.session.add(user)
        return True

    @staticmethod
    def verify_access_token(access_token, refresh_token=None):
        token = Token.query.filter(Token.access_token == access_token).first()
        if token is not None:
            if token.access_expiration > datetime.utcnow():
                token.user.ping()
                db.session.commit()
                if token.user.role.name != 'System user':
                    return token.user

    @staticmethod
    def verify_refresh_token(refresh_token, access_token):
        token = Token.query.filter((Token.refresh_token == refresh_token) & (Token.access_token == access_token)).first()
        if token is not None:
            if token.refresh_expiration > datetime.utcnow():
                return token
            # someone tried to refresh with an expired token
            # revoke all tokens from this user as a precaution
            token.user.revoke_auth_tokens()
            db.session.commit()

    def can(self, permission):
        return self.role.has_permission(permission)

    def confirm(self, confirmation_token):
        try:
            payload = jwt.decode(
                confirmation_token,
                current_app.config['SECRET_KEY'],
                algorithms=['HS256'],
                issuer=current_app.config['SERVER_NAME'],
                options={'require': ['exp', 'iat', 'iss', 'purpose', 'sub']}
            )
            current_app.logger.warning(payload)
        except jwt.PyJWTError:
            return False
        if payload.get('purpose') != 'user.confirm':
            return False
        if payload.get('sub') != self.hashid:
            return False
        self.confirmed = True
        db.session.add(self)
        return True

    def delete(self):
        shutil.rmtree(self.path, ignore_errors=True)
        db.session.delete(self)

    def generate_auth_token(self):
        return Token(
            access_token=secrets.token_urlsafe(),
            access_expiration=datetime.utcnow() + timedelta(minutes=15),
            refresh_token=secrets.token_urlsafe(),
            refresh_expiration=datetime.utcnow() + timedelta(days=7),
            user=self
        )

    def generate_confirm_token(self, expiration=3600):
        now = datetime.utcnow()
        payload = {
            'exp': now + timedelta(seconds=expiration),
            'iat': now,
            'iss': current_app.config['SERVER_NAME'],
            'purpose': 'user.confirm',
            'sub': self.hashid
        }
        return jwt.encode(
            payload,
            current_app.config['SECRET_KEY'],
            algorithm='HS256'
        )

    def generate_reset_password_token(self, expiration=3600):
        now = datetime.utcnow()
        payload = {
            'exp': now + timedelta(seconds=expiration),
            'iat': now,
            'iss': current_app.config['SERVER_NAME'],
            'purpose': 'User.reset_password',
            'sub': self.hashid
        }
        return jwt.encode(
            payload,
            current_app.config['SECRET_KEY'],
            algorithm='HS256'
        )

    def is_administrator(self):
        return self.can(Permission.ADMINISTRATE)

    def ping(self):
        self.last_seen = datetime.utcnow()

    def revoke_auth_tokens(self):
        for token in self.tokens:
            db.session.delete(token)

    def verify_password(self, password):
        if self.role.name == 'System user':
            return False
        return check_password_hash(self.password_hash, password)

    #region Profile Privacy settings
    def has_profile_privacy_setting(self, setting):
        return self.profile_privacy_settings & setting == setting
    
    def add_profile_privacy_setting(self, setting):
        if not self.has_profile_privacy_setting(setting):
            self.profile_privacy_settings += setting

    def remove_profile_privacy_setting(self, setting):
        if self.has_profile_privacy_setting(setting):
            self.profile_privacy_settings -= setting

    def reset_profile_privacy_settings(self):
        self.profile_privacy_settings = 0
    #endregion Profile Privacy settings

    def follow_corpus(self, corpus):
        if not self.is_following_corpus(corpus):
            self.followed_corpora.append(corpus)

    def unfollow_corpus(self, corpus):
        if self.is_following_corpus(corpus):
            self.followed_corpora.remove(corpus)

    def is_following_corpus(self, corpus):
        return corpus in self.followed_corpora


    def to_json_serializeable(self, backrefs=False, relationships=False, filter_by_privacy_settings=False):
        json_serializeable = {
            'id': self.hashid,
            'confirmed': self.confirmed,
            'email': self.email,
            'last_seen': (
                None if self.last_seen is None
                else self.last_seen.strftime('%Y-%m-%d %H:%M')
            ),
            'member_since': self.member_since.strftime('%Y-%m-%d'),
            'username': self.username,
            'full_name': self.full_name,
            'about_me': self.about_me,
            'website': self.website,
            'location': self.location,
            'organization': self.organization,
            'job_status_mail_notification_level': \
                    self.setting_job_status_mail_notification_level.name,
            'is_public': self.is_public,
            'show_email': self.has_profile_privacy_setting(ProfilePrivacySettings.SHOW_EMAIL),
            'show_last_seen': self.has_profile_privacy_setting(ProfilePrivacySettings.SHOW_LAST_SEEN),
            'show_member_since': self.has_profile_privacy_setting(ProfilePrivacySettings.SHOW_MEMBER_SINCE)
        }
        json_serializeable['avatar'] = (
            None if self.avatar is None
            else self.avatar.to_json_serializeable(relationships=True)
        )
        if backrefs:
            json_serializeable['role'] = \
                self.role.to_json_serializeable(backrefs=True)
        if relationships:
            json_serializeable['corpus_follower_associations'] = {
                x.hashid: x.to_json_serializeable(relationships=True)
                for x in self.corpus_follower_associations
            }
            json_serializeable['corpora'] = {
                x.hashid: x.to_json_serializeable(relationships=True)
                for x in self.corpora
            }
            json_serializeable['jobs'] = {
                x.hashid: x.to_json_serializeable(relationships=True)
                for x in self.jobs
            }
            json_serializeable['tesseract_ocr_pipeline_models'] = {
                x.hashid: x.to_json_serializeable(relationships=True)
                for x in self.tesseract_ocr_pipeline_models
            }
            json_serializeable['spacy_nlp_pipeline_models'] = {
                x.hashid: x.to_json_serializeable(relationships=True)
                for x in self.spacy_nlp_pipeline_models
            }

        if filter_by_privacy_settings:
            if not self.has_profile_privacy_setting(ProfilePrivacySettings.SHOW_EMAIL):
                json_serializeable.pop('email')
            if not self.has_profile_privacy_setting(ProfilePrivacySettings.SHOW_LAST_SEEN):
                json_serializeable.pop('last_seen')
            if not self.has_profile_privacy_setting(ProfilePrivacySettings.SHOW_MEMBER_SINCE):
                json_serializeable.pop('member_since')
        return json_serializeable


class TesseractOCRPipelineModel(FileMixin, HashidMixin, db.Model):
    __tablename__ = 'tesseract_ocr_pipeline_models'
    # Primary key
    id = db.Column(db.Integer, primary_key=True)
    # Foreign keys
    user_id = db.Column(db.Integer, db.ForeignKey('users.id'))
    # Fields
    title = db.Column(db.String(64))
    description = db.Column(db.String(255))
    version = db.Column(db.String(16))
    compatible_service_versions = db.Column(ContainerColumn(list, 255))
    publisher = db.Column(db.String(128))
    publisher_url = db.Column(db.String(512))
    publishing_url = db.Column(db.String(512))
    publishing_year = db.Column(db.Integer)
    is_public = db.Column(db.Boolean, default=False)
    # Relationships
    user = db.relationship('User', back_populates='tesseract_ocr_pipeline_models')

    @property
    def path(self):
        return os.path.join(
            self.user.path,
            'tesseract_ocr_pipeline_models',
            str(self.id)
        )

    @property
    def jsonpatch_path(self):
        return f'{self.user.jsonpatch_path}/tesseract_ocr_pipeline_models/{self.hashid}'

    @property
    def url(self):
        return url_for(
            'contributions.tesseract_ocr_pipeline_model',
            tesseract_ocr_pipeline_model_id=self.id
        )

    @property
    def user_hashid(self):
        return self.user.hashid

    @staticmethod
    def insert_defaults():
        nopaque_user = User.query.filter_by(username='nopaque').first()
        defaults_file = os.path.join(
            os.path.dirname(os.path.abspath(__file__)),
            'TesseractOCRPipelineModel.defaults.yml'
        )
        with open(defaults_file, 'r') as f:
            defaults = yaml.safe_load(f)
        for m in defaults:
            model = TesseractOCRPipelineModel.query.filter_by(title=m['title'], version=m['version']).first()  # noqa
            if model is not None:
                model.compatible_service_versions = m['compatible_service_versions']
                model.description = m['description']
                model.publisher = m['publisher']
                model.publisher_url = m['publisher_url']
                model.publishing_url = m['publishing_url']
                model.publishing_year = m['publishing_year']
                model.is_public = True
                model.title = m['title']
                model.version = m['version']
                continue
            model = TesseractOCRPipelineModel(
                compatible_service_versions=m['compatible_service_versions'],
                description=m['description'],
                publisher=m['publisher'],
                publisher_url=m['publisher_url'],
                publishing_url=m['publishing_url'],
                publishing_year=m['publishing_year'],
                is_public=True,
                title=m['title'],
                user=nopaque_user,
                version=m['version']
            )
            db.session.add(model)
            db.session.flush(objects=[model])
            db.session.refresh(model)
            model.filename = f'{model.id}.traineddata'
            r = requests.get(m['url'], stream=True)
            pbar = tqdm(
                desc=f'{model.title} ({model.filename})',
                unit="B",
                unit_scale=True,
                unit_divisor=1024,
                total=int(r.headers['Content-Length'])
            )
            pbar.clear()
            with open(model.path, 'wb') as f:
                for chunk in r.iter_content(chunk_size=1024):
                    if chunk:  # filter out keep-alive new chunks
                        pbar.update(len(chunk))
                        f.write(chunk)
                pbar.close()
        db.session.commit()

    def delete(self):
        try:
            os.remove(self.path)
        except OSError as e:
            current_app.logger.error(e)
        db.session.delete(self)

    def to_json_serializeable(self, backrefs=False, relationships=False):
        json_serializeable = {
            'id': self.hashid,
            'compatible_service_versions': self.compatible_service_versions,
            'description': self.description,
            'publisher': self.publisher,
            'publisher_url': self.publisher_url,
            'publishing_url': self.publishing_url,
            'publishing_year': self.publishing_year,
            'is_public': self.is_public,
            'title': self.title,
            'version': self.version,
            **self.file_mixin_to_json_serializeable()
        }
        if backrefs:
            json_serializeable['user'] = \
                self.user.to_json_serializeable(backrefs=True)
        return json_serializeable


class SpaCyNLPPipelineModel(FileMixin, HashidMixin, db.Model):
    __tablename__ = 'spacy_nlp_pipeline_models'
    # Primary key
    id = db.Column(db.Integer, primary_key=True)
    # Foreign keys
    user_id = db.Column(db.Integer, db.ForeignKey('users.id'))
    # Fields
    title = db.Column(db.String(64))
    description = db.Column(db.String(255))
    version = db.Column(db.String(16))
    compatible_service_versions = db.Column(ContainerColumn(list, 255))
    publisher = db.Column(db.String(128))
    publisher_url = db.Column(db.String(512))
    publishing_url = db.Column(db.String(512))
    publishing_year = db.Column(db.Integer)
    pipeline_name = db.Column(db.String(64))
    is_public = db.Column(db.Boolean, default=False)
    # Relationships
    user = db.relationship('User', back_populates='spacy_nlp_pipeline_models')

    @property
    def path(self):
        return os.path.join(
            self.user.path,
            'spacy_nlp_pipeline_models',
            str(self.id)
        )

    @property
    def jsonpatch_path(self):
        return f'{self.user.jsonpatch_path}/spacy_nlp_pipeline_models/{self.hashid}'

    @property
    def url(self):
        return url_for(
            'contributions.spacy_nlp_pipeline_model',
            spacy_nlp_pipeline_model_id=self.id
        )

    @property
    def user_hashid(self):
        return self.user.hashid

    @staticmethod
    def insert_defaults():
        nopaque_user = User.query.filter_by(username='nopaque').first()
        defaults_file = os.path.join(
            os.path.dirname(os.path.abspath(__file__)),
            'SpaCyNLPPipelineModel.defaults.yml'
        )
        with open(defaults_file, 'r') as f:
            defaults = yaml.safe_load(f)
        for m in defaults:
            model = SpaCyNLPPipelineModel.query.filter_by(title=m['title'], version=m['version']).first()  # noqa
            if model is not None:
                model.compatible_service_versions = m['compatible_service_versions']
                model.description = m['description']
                model.publisher = m['publisher']
                model.publisher_url = m['publisher_url']
                model.publishing_url = m['publishing_url']
                model.publishing_year = m['publishing_year']
                model.is_public = True
                model.title = m['title']
                model.version = m['version']
                model.pipeline_name = m['pipeline_name']
                continue
            model = SpaCyNLPPipelineModel(
                compatible_service_versions=m['compatible_service_versions'],
                description=m['description'],
                publisher=m['publisher'],
                publisher_url=m['publisher_url'],
                publishing_url=m['publishing_url'],
                publishing_year=m['publishing_year'],
                is_public=True,
                title=m['title'],
                user=nopaque_user,
                version=m['version'],
                pipeline_name=m['pipeline_name']
            )
            db.session.add(model)
            db.session.flush(objects=[model])
            db.session.refresh(model)
            model.filename = m['url'].split('/')[-1]
            r = requests.get(m['url'], stream=True)
            pbar = tqdm(
                desc=f'{model.title} ({model.filename})',
                unit="B",
                unit_scale=True,
                unit_divisor=1024,
                total=int(r.headers['Content-Length'])
            )
            pbar.clear()
            with open(model.path, 'wb') as f:
                for chunk in r.iter_content(chunk_size=1024):
                    if chunk:  # filter out keep-alive new chunks
                        pbar.update(len(chunk))
                        f.write(chunk)
                pbar.close()
        db.session.commit()
    
    def delete(self):
        try:
            os.remove(self.path)
        except OSError as e:
            current_app.logger.error(e)
        db.session.delete(self)

    def to_json_serializeable(self, backrefs=False, relationships=False):
        json_serializeable = {
            'id': self.hashid,
            'compatible_service_versions': self.compatible_service_versions,
            'description': self.description,
            'publisher': self.publisher,
            'publisher_url': self.publisher_url,
            'publishing_url': self.publishing_url,
            'publishing_year': self.publishing_year,
            'pipeline_name': self.pipeline_name,
            'is_public': self.is_public,
            'title': self.title,
            'version': self.version,
            **self.file_mixin_to_json_serializeable()
        }
        if backrefs:
            json_serializeable['user'] = self.user.to_json_serializeable(backrefs=True)
        return json_serializeable


class JobInput(FileMixin, HashidMixin, db.Model):
    __tablename__ = 'job_inputs'
    # Primary key
    id = db.Column(db.Integer, primary_key=True)
    # Foreign keys
    job_id = db.Column(db.Integer, db.ForeignKey('jobs.id'))
    # Relationships
    job = db.relationship(
        'Job',
        back_populates='inputs'
    )

    def __repr__(self):
        return f'<JobInput {self.filename}>'

    @property
    def content_url(self):
        return url_for(
            'jobs.download_job_input',
            job_id=self.job.id,
            job_input_id=self.id
        )

    @property
    def jsonpatch_path(self):
        return f'{self.job.jsonpatch_path}/inputs/{self.hashid}'

    @property
    def path(self):
        return os.path.join(self.job.path, 'inputs', str(self.id))

    @property
    def url(self):
        return url_for(
            'jobs.job',
            job_id=self.job_id,
            _anchor=f'job-{self.job.hashid}-input-{self.hashid}'
        )

    @property
    def user_hashid(self):
        return self.job.user.hashid

    @property
    def user_id(self):
        return self.job.user.id

    def to_json_serializeable(self, backrefs=False, relationships=False):
        json_serializeable = {
            'id': self.hashid,
            **self.file_mixin_to_json_serializeable()
        }
        if backrefs:
            json_serializeable['job'] = \
                self.job.to_json_serializeable(backrefs=True)
        return json_serializeable


class JobResult(FileMixin, HashidMixin, db.Model):
    __tablename__ = 'job_results'
    # Primary key
    id = db.Column(db.Integer, primary_key=True)
    # Foreign keys
    job_id = db.Column(db.Integer, db.ForeignKey('jobs.id'))